after seeing the amount of json/toml files being posted in the last few threads it's probably very confusing for anyone trying to start learning SDXL lora training.
So I decided to gather the popular ones and train them all with my old raita dataset.
I know training settings will vary between different datasets but I can at least test which one looks more faithful to the artist at first glance.
The raita loras aren't perfect as thier just simple tests for now sodon't expect anything good for now
I've also added the artists NDC/ponkotsuu to the mega
Anonymous 02/17/24(Sat)10:50:10 No.7825805
catbox_6gy4ea.png (2.71 MB, 1432x1840)
>>7825790
It's my own
https://mega.nz/folder/e6xzFZbJ#ZZl6dlUQkaTeyzG_4OpJtg
after seeing the amount of json/toml files being posted in the last few threads it's probably very confusing for anyone trying to start learning SDXL lora training.
So I decided to gather the popular ones and train them all with my old raita dataset.
I know training settings will vary between different datasets but I can at least test which one looks more faithful to the artist at first glance.
The raita loras aren't perfect as thier just simple tests for now sodon't expect anything good for now
I've also added the artists NDC/ponkotsuu to the mega